get_mlflow_host_token() does not load token in environment variable.

[SDonkelaarGDD]

We want to use mlflow-export-import to migrate models between OOS tracking servers in an enterprise setting (at a bank). However, since our tracking servers are both behind oauth2 proxies, support for bearer tokens is essential for us to make it work.

I inspected the code and the reason for this is that the function def get_mlflow_host_token() does not actually loads the token from the environment variable, and hence returns none.

Is this on purpose? Otherwise., I created a PR that fixes the issue.

[SDonkelaarGDD]

@amesar Gentle reminder. This issue is blocking us from using it and therefore we're required to use a fork of the repo.

[bin0110001]

I've been having the same issue and was considering making the exact same pull request I was going to. However, I did some more digging and might have an alternative solution that might work for you (though I would still hope they either accept the pull request or update documentation). If you set the MLFLOW_TRACKING_URI to a static string of "databricks" it will return the environment variables for DATABRICKS_HOST and DATABRICKS_TOKEN, assuming you don't have a spark configuration or .databrickscfg file and you don't have the "MLFLOW_PROFILE" environment variable. Ignore the comments on get_host_token that only mention the databrickscfg file as the default provider appears to iterate over the 3 different sources.

[davehowell]

@bin0110001 I found a similar workaround, using the databrickscfg file works if you name the profile [databricks] rather than [DEFAULT]